extract lumped model from S parameter
I am modeling an on chip inductor and balun. The S parameter file has been generated from EM simulation.
Now, I am confused with extract lumped model from this S parameter File. ADS optimize tool was used to optimize the lumped model, it is ok with the inductor, but result in a large error with the balun.
how do you guys extract lumped model form S parameter, which tools can do this thing? Is there any refernce on doing this thing?
Thanks
hhxuexia
Please observe forum rules related to cross-posting.
A model is a simplification, in case of a distributed system it's also involving principle limitations. Wideband S parameter representations of a device don't have a naturally corresponding lumped model. It's up to the experts intelligency to find an approriate one.
For specific problems, you'll find suggestions in literature.
S parameter measurement of a balun is 3- or 4-port problem (requires differential + common mode measurements at least for the balanced side). How did you consider this properties in your model?
That's the usual procedure today: take a suitable lumped element model from literature, and then calculate/fit the component values.
At Sonnet, we are just releasing a new version (13.56) that has automated lumped model extraction for two-port inductors on silicion (without center tap). The next step will be to automate the extraction of center tapped inductors. The difficulty is that every topology needs to be developed and implemented separately.
In the past, Sonnet had already implemented a broadband SPICE model extraction, which will automatically fit anything, with arbitrary topology, but the result of that is a "black block" with controlled sources, which can not be physically interpreted. Customers use it for inductors and transfomers, and it works fine, but many customers would like a simple, fixed model that can be physically interpreted.
A recent paper on methodology for extracting inductor lumped models from S/Y/Z-parameters:
http://ieeexplore.ieee.org/xpl/freea...number=6107596